Now, although all this is to move in a very uncertain ambit, harassed by a multitude of diverse and vain dilettantisms and mysticisms, and only too frequently by fraud, it is not any longer possible nowadays to deny that facts, objectively known, compel the positive scientist to have recourse to some such suppositions. Also without making the "subliminal," with Myers, a kind of "deus ex machina" in the world, it is certain that mediumistic phenomena of the kind mentioned are henceforth to be considered as a subject of study for an open-minded psychology. I may refer in support of this view, among others, to the powerful work of Morselli. And to return to the "thinking" animals, we find that the mediumistic hypothesis, however shifty it may seem, is a better explanation than the telepathic hypothesis--which has already itself become rather more systematized in modern psychology. After his visits to Elberfeld, Claparede, as I said, had found it difficult to treat as valid the telepathic hypothesis when applied to Krall's horses. What, indeed, had been "transmitted" to them? Numbers? Words? Single letters? (or orders to stop the foot at the right time?) It must be remembered that the horses were tapping their answers by using a sort of stenography, that usually left out the vowels: that besides, although the words could be recognized in the most certain manner, the spelling was most irregular, and, as I have already pointed out, sometimes reversed. Further, as to the words themselves, most infantile phrases were used, certainly such as no adult would have suggested. Was it suggestion then from one unconscious to another? But this is to fall back upon a supposition of the "mediumistic" type, and takes no count of the cases of replies to questions which were unknown to everybody present, and brings us to the single dilemma: either there is intelligence in the human sense in the animal, or a relationship of the mediumistic type above described between the several minds concerned. As to the interesting observations reported by Ferrari and Pulle, it seems to me opportune to quote here some extracts from the first of these distinguished authors. "This seance was particularly interesting, because I find it recorded in my notes that a fact was verified three times consecutively, which had occurred sporadically more than once before, and had been observed and noted by us and various other witnesses. "It consisted in this: While I was putting in the box the number of balls which I had intended the horse to read, the horse, which often could not even have seen the number of balls, because I covered them partly with my head and hands, tapped out the correct number.